Kinect-based Mobile Robot Navigation System

The key point in autonomous mobile navigation robot system is localization and mapping. Scanning sensors that these systems use are mostly based on laser. As lasers are expensive and they can just be used to create two-dimensional map, we use XBOX360 Kinect provided by Microsoft as the vision sensor in our navigation system. This paper proposes a method of using Kinect to build a mobile robot navigation system as a sensor. Navigation system mainly includes three modules: based on the closed-loop detection algorithm of the SLAM technology to create a 2D environment map, based on the improved D* algorithm of robot path planning in dynamic environment, The design of navigation system based on Kinect. The main work is as follows:(1) First of all, based on the RGB-D image, we propose a closed-loop detection 2DSLAM algorithm. the RGB-D image of Kinect is obtained by using SURF algorithm for feature extraction and matching, and then use the RANSAC(random sample consensus algorithm to remove the false points) possible matches to complete the initial registration of RGB image, by using the generalized ICP(iterative closest) estimation and optimization algorithm to further optimize the initial registration in order to complete the feature point position.(2) Secondly, how the robot path planning in dynamic environment based on effectively, the propose a kind of improved D* algorithm based on the robot dynamic path planning, using C++ to improve D* algorithm were targeted to achieve that, in under the limited cost through experiment, calculation ability and hardware equipment will be able to achieve very good path planning in complex environment, in order to verify the feasibility of its application in certain conditions.(3) Finally, the experimental platform for the study and result, including the experimental platform of the TurtleBot, robot operating system and environment sensing camera Kinect introduction. The use of hardware design of TurtleBot robot, a navigation system to complete its autonomous navigation capabilities, through the specific experiment has proved that this scheme is feasible.